0
\$\begingroup\$

I have a project I am designing whose fundamental parts consist of sensors connected to a micro-controller. the micro-controller will then send the data out over an ethernet port using standard tcp/ip protocols. I have already created the a proof of concept for the project using an Arduino Yun, but ran into several issues when the libraries and apis that make the Arduino easy to use seem to have prevented me from directly manipulating the registers that control the Arduino's ADC. Additionally, designing the physical ethernet circuitry would allow me to be more flexible with the board layout and housing design and allow me to choose micro-controllers without built in ethernet circuitry if I wish. However, I am having trouble designing the ethernet hardware due to my limited experience.

It would help me greatly if there were circuit designs I could use as a reference when designing my circuit. Therefore my question is does the 802.3 Ethernet standard have a set of recommended circuit designs in schematic and/or block diagram form or does it solely lay out the features and functions for the standard? It would also be a helpful if someone could point out some surface mount ICs commonly used in mass manufacture Ethernet socket designs.

\$\endgroup\$
7
  • \$\begingroup\$ UART? Most microcontrollers with Ethernet will have some reference circuitry, but it depends whether you have integrated PHY. See electronics.stackexchange.com/questions/297/… \$\endgroup\$
    – pjc50
    Commented Dec 19, 2015 at 10:49
  • \$\begingroup\$ PHY is the physical hardware correct? I am unfamiliar with the term. \$\endgroup\$
    – Josh Nabours
    Commented Dec 19, 2015 at 12:08
  • \$\begingroup\$ I just read the article you linked to. I didn't realize the OSI physical layer could be abbreviated to PHY or that there were micro-controllers with integrated Ethernet Hardware. I have only used Freescale's micro-controller so far. \$\endgroup\$
    – Josh Nabours
    Commented Dec 19, 2015 at 12:17
  • \$\begingroup\$ The PHY is the implementation of a particular ethernet layer(e.g. 10/100tx) typically all a PHY needs is just magnetics (isolation) and an ether net jack for 10/100TX The suggested circuit and supporting components should be in datasheet for the PHY or mcu (w integrated phy). Typically a discrete PHY connects to an mcu through a MII or related interface (formally defined as a layer 1 sublayer) \$\endgroup\$
    – crasic
    Commented Dec 19, 2015 at 17:09
  • \$\begingroup\$ I see you mentioned 10/100Tx Ethernet but not gigabit ethernet. Are micro-controllers typically not capable of supporting gigabit ethernet because of a lack of a 1GHz clock signal to drive the data transfer functionality? \$\endgroup\$
    – Josh Nabours
    Commented Dec 19, 2015 at 17:35

1 Answer 1

Reset to default
0
\$\begingroup\$

IEEE802.3 is thousands of pages covering everything from 1 Mb/s to 100 Gb/s systems. You can download and read the standard for yourself here.

IEEE provides little in the way of reference implementations. However the chipmakers who provide the PHY devices will typically provide reference designs. These designs will, however, depend on you to provide a uC implementing all the higher layers of the ethernet stack (from the MAC on up).

\$\endgroup\$
6
  • \$\begingroup\$ I expected it to be big (the usb specs are 600 something pages) but I didn't think it would be more than 1000. That means this undertaking would be similar in scale and scope to designing your own add on network card for a PC if you don't use a micro-controller with built in ethernet huh? I appreciate you answering my question. Now I can go digging through app notes instead of searching the 802.3 standard. Though I will end up reading that too. \$\endgroup\$
    – Josh Nabours
    Commented Dec 19, 2015 at 17:47
  • \$\begingroup\$ If you narrow down your need (like, say you're only interested in 100 Mb/s, not gigabit ethernet, 10G, 100G, etc), it'll reduce a lot what you need to read. But even knowing what to read is complicated; and the language used to describe the requirements is not obvious if you don't already know what it means. \$\endgroup\$
    – The Photon
    Commented Dec 19, 2015 at 17:48
  • \$\begingroup\$ I figured as much. That is part of why I asked the question too because I need to better define what my project actually requires so I can pick the cheapest possible hardware parts that get the job done properly. \$\endgroup\$
    – Josh Nabours
    Commented Dec 19, 2015 at 17:52
  • \$\begingroup\$ The spec is more or less useless to you. Unless you are designing your own stack, MAC, and/or PHY from scratch. What you need is to research MCUs, MAC, and PHY components and look at reference schematics in datasheets and application notes from these part manufacturers. You can pick the arduino ethernet shield as a jumping off point \$\endgroup\$
    – crasic
    Commented Dec 19, 2015 at 18:21
  • \$\begingroup\$ That being said I would read the spec for due diligence and formal knowledge once you've identified what part of it is relevant to you \$\endgroup\$
    – crasic
    Commented Dec 19, 2015 at 18:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.